I have the following situation here: our native work environment till now was linux, however due to a need to comply to wider corporate interoperation, integration and more effective communication we will need to install windows on our machines.
Having two OS'es is not a solution for me, I don't want to reboot the computer every time I have to join a meeting or anything else windows related, or the other way around reboot after a meeting to start the real work. Till now, I managed to get along the "inverted" way (vmware with windows inside the linux box) however now I'd like to try out the other way: vmware running on windows, with a linux as development environment, windows is being kept only for meetings, mail, etc.
Our solution will still be linux based and I will do the main programming stuff (and compiling) inside the vmware... but the following problems come around in this case:
our linux application requires access to hardware (capture cards, such as BlackMagic, DVS Centaurus, BlueFish, etc...). This was not an issue till now, since we had linux as the main OS with direct access to the cards and after some careful fine tuning the cards were working nicely... However, now that we will have this "thin layer" of virtualization between our app and the physical hardware I am curious if anyone of you ever managed (or knows for sure that it's not possible) to make a card like this to work under vmware?
our application is mainly written in c/c++ ... anyone of you has any experience if committing to compilation under a vmware will significantly increase the compile time? Right now a full rebuild takes around 15 minutes for the whole app, but on a damn powerful machine (8 cores, 12GB of RAM).
our application is pretty CPU hungry... it does some video encoding/decoding realtime. This was not an issue till now (on the machine specified above). However, has anyone of you any experience about the speed reduction of a real-time application when moved to vmware? And the same question applies to the network layer too. We send and receive a lot of data.
Thank you for your time